Draft version February 15, 2010Preprint typeset using LATEX style emulateapj v. 11/10/09

THE CERRO CHAJNANTOR ATACAMA TELESCOPE (CCAT)

Michel Fich, Pauline Barmby, Pierre Bastien, Jan Cami, Stephane Claude, James Di Francesco, SarahGallagher, Mark Halpern, Martin Houde, Doug Johnstone, Gilles Joncas, Brenda Matthews, Els Peeters,

Rene Plume, Russell Redman, Michael Reid, Gerald Schieven, Douglas Scott, Tracy Webb and theInternational CCAT Team

Subject headings:

1. SUMMARY

“How did we get from energy fluctuations in the earlyUniverse to galaxies, stars and planets?”. This is oneof the fundamental problems in astrophysics, and the 25meter diameter CCAT is a unique project geared towardsthe investigation of this question. CCAT will have a sen-sitivity similar to that of ALMA (the Atacama LargeMillimeter Array) at submillimeter wavelengths but willbe much faster than ALMA when mapping in the con-tinuum. Located on the peak of Cerro Chajnantor closeto, but significantly higher, than the ALMA site, will al-low CCAT to operate at wavelengths as short as 200 µm,considerably better than the ALMA wavelength limit.

New technologies are allowing us to build submillime-ter/far infrared cameras of unprecedented size, fully pop-ulating the focal plane of the telescope. CCAT will havea large enough field of view to accommodate a detector 20arcminutes in diameter and populated with greater than100 kilopixels. When this larger and even more sensitivecamera is combined with the excellent site, CCAT willmap the sky 1000 times faster and with better resolu-tion than SCUBA-2, the current state-of-the-art detec-tor. CCAT will also be equipped with a variety of otherinstruments including spectrometers such as broad-banddirect-detection grating instruments (such as the existingZEUS and Z-spec at the CSO), heterodyne focal planearrays (such as HARP at the JCMT), and Fourier Trans-form Spectrometers.

CCAT is currently a joint project between CornellUniversity, the California Institute of Technology, JetPropulsion Laboratory, the University of Colorado, arapidly growing consortium of Canadian universities, theUnited Kingdom Science and Technology Facilities Coun-cil (STFC), and the Universities of Bonn & Koln.

CCAT will arrive at a time when ALMA will havebeen in use for at least five years and no other compa-rable submm facility is likely to exist (with both JCMTand Herschel having reached the end of their operationsat least four years before CCAT can begin operations).Canadian contributions to CCAT will likely include theobservatory dome, parts of the continuum cameras, het-erodyne focal plane arrays, FTS, polarimeter, and soft-ware for both data analysis and for the data archive.Thus, CCAT will continue the legacy of the JCMT andguarantee that a strong Canadian academic traditioncontinues to flourish in the increasingly important areaof submm science, training, and instrumentation.

2. KEY SCIENCE GOALS

In this Section, due to space restrictions we briefly dis-cuss a only few of the science paths CCAT will tread.Broader and more detailed descriptions can be foundin Science White Papers by Scott et al., Matthews etal.,Johnstone et al., and Barmby et al..

2.1. Measuring the Star Formation History of GalaxiesAcross Cosmic Time

A comprehensive picture of galaxy formation and evo-lution must account for the bolometric luminosities ofgalaxies across cosmic time. COBE revealed a cosmic far-infrared (FIR) extragalactic background radiation withflux approximately equal to the integrated extragalac-tic optical and ultraviolet starlight in the Universe. Themost luminous star forming galaxies emit the bulk oftheir light in the FIR/submm and most of the lightfrom high z galaxies reaches us in that spectral domain.Submm spectral probes are keenly sensitive to physicalconditions of the gas, thereby elucidating the context forstar formation and providing a crucial observational linkbetween the buildup of the stellar masses and centralsupermassive black holes.

To determine the amount of energy that has been re-leased by galaxies, it is essential to measure their rest-frame FIR radiation, which peaks at 50 - 200 µm andis redshifted into the submm bands for z > 1. Dustemission comprises 50% of the total integrated luminos-ity of galaxies. That fraction is even larger for the mostluminous galaxies and galaxies at high z. The amountof FIR/submm emission – and thus of star formation –from dusty galaxies is impossible to infer from the spec-tral properties of the escaping optical/UV light alone.Submm observations have a strong advantage in searchesfor high redshift galaxies: the negative K-correction. Be-cause the slope of the product of the Planck functionand the emissivity function of dust grains varies steeplywith frequency on the Rayleigh-Jeans side of the spec-trum (S! ! !3!4), the observed brightness of a galaxyis independent of z from 1 < z < 10, in that spectralregime. Thus submm surveys provide a natural meansfor identifying high-redshift (z > 5) galaxy candidateswithin large-scale surveys: those with weak 200 - 700µm emission and bright 800 µm to mm–wave emission.

SCUBA-2 on the JCMT and the Herschel Space Ob-servatory will have an important impact in this fieldin the next few years. However the confusion limit forCCAT will be significantly better than for either of thesetelescopes allowing CCAT to reach substantially furtherdown the luminosity function. The greatest gain for

2

CCAT over SCUBA-2 may be due to its smaller beamat 200 and 350 µm then SCUBA-2 achieves at 450 µm.The CCAT resolution will be good enough that reliablematches to optical and near-IR surveys will be possible,something that has proven very di!cult at the resolutionthat the JCMT provides.

Furthermore, Herschel’s cryogen will be exhausted ap-proximately when ALMA comes on line, precluding anyopportunities for coordinated observations or surveys.On CCAT, a 10 square-degree survey – covering a cosmo-logically relevant volume – could be covered to a depth of0.2 mJy rms at 350 µm in 2,000 hours, yielding of order105 mostly faint, distant galaxies. Moreover, CCAT willcarry out a spectroscopic survey of submm galaxies, us-ing multi–object versions of broadband direct-detectiongrating spectrometers such as Z–Spec and ZEUS, now inuse at the CSO.

2.2. Star Formation and the ISM in Nearby GalaxiesTo understand the strong evolution of star formation

over cosmic time, nearby, spatially resolved galaxies mustbe studied to relate the astrophysical probes to high zsystems: resolved submm images will reveal the interplaybetween the star formation process and the natal inter-stellar medium (ISM). Of particular interest are the mostactive regions in nearby normal and starburst galaxies;these regions will provide the best templates for the dis-tant, LIRG and ULIRG-class galaxies responsible for thebulk of the cosmic FIR background.

Multi-wavelength studies in concert with submm ob-servations will address fundamental questions about starformation in galaxies, such as: What are the relation-ships between the age (chemical abundances) of the ISM,the degree of star formation activity, galactic morphol-ogy, and the environment? What triggers galaxy-widestarbursts? Do starbursts end by consuming all the avail-able fuel or by disrupting the natal environment throughstellar winds?

Multi-band images can trace the process of gas com-pression in spiral density waves, the formation of starsin molecular cloud cores, and the disruption of the par-ent clouds by newly formed stars. The FIR/submmspectral regime provides a wide variety of extinction-freespectral–line probes of both ambient radiation fields andthe physical properties of interstellar gas (e.g., density,temperature, dynamics, radiation intensity and hard-ness). Most of those lines lie within a few hundred K ofthe ground state and have modest critical densities; theemitted radiation is nearly always optically thin. Theyare important (often dominant) coolants for the phasesof the ISM relevant to star formation processes.

The most important FIR and submm lines include fine-structure lines from abundant species (C, CII, NII, NIII,OI, OIII), plus the J = 4–3 to 13–12 rotational transi-tions of CO. These lines are very bright in star forminggalaxies, often summing to more than 1% of the totalgalaxy luminosity. With Fabry-Perot, Fourier transform,or waveguide–fed multi-object spectrometers, CCAT willbe able to deliver spectroscopic images of nearby galax-ies in the [NII] 205 µm, [CI] 370 and 609 µm, mid–J CO(e.g. 4–3, 6–5, 7–6) and 13CO(6–5 and 8–7) rotationallines at angular resolutions as fine as 2"". In the nuclei ofsome galaxies (e.g. ULIRGs), it would detect CO emis-sion up to J = 13-12 (200 µm) arising from nuclear clouds

highly excited by starbursts, or even CO emission fromAGN-excited molecular tori, thus providing a link be-tween stellar mass buildup and supermassive black holegrowth.

2.3. The Origin of the Stellar IMFIn the Milky Way, it is often claimed that the stel-

lar Initial Mass Function (IMF) is remarkably consistentacross a variety of environments. The origin of this uni-formity, if true, is unknown. Among the physical pro-cesses that may lead to a seemingly invariant IMF aregravitational or turbulent fragmentation, feedback fromstellar winds and outflows, competitive accretion, ejec-tion of protostellar cores and stellar mergers.

An intriguing possibility is that the mass function ofdense clumps in molecular clouds, identified by theirthermal dust emission, has a similar shape to the stellarIMF, suggesting that the clump mass function translatesinto the stellar IMF. CCAT observations will establish ifthe clump mass function follows the stellar IMF to thesubstellar (brown dwarf) regime, and if the clump massfunction is similar over a wide range of environments inthe Galaxy. If the clump mass function is invariant, itwill provide compelling evidence that the stellar IMF isimprinted in the fragmentation structure of molecularclouds.

Existing measures of the clump mass function arevery heterogeneous and therefore hard to intercompare.CCAT will be able to measure the mass function betterthan any other instrument including the largest scales,di!cult for an interferometer such as ALMA, and done tosubstellar scales where the resolution of existing single-dish instruments is not good enough. Sensitivity, res-olution, mapping speed and "–coverage of CCAT willuniquely enable Galactic surveys that can link the stel-lar IMF to the physics and topology of the ISM.

2.4. The Sizes of Trans–Neptunian ObjectsIn recent years, several hundred Solar System objects

beyond Neptune have been discovered (TNOs). Theyare believed to have formed in the outer reaches of theprotoplanetary disk around the Sun and to have under-gone very little evolution since their early formation. Theprimitive nature of the material in this region holds im-portant clues towards our understanding of the formationand evolution of the Solar System. The measurement ofTNO sizes is important not only to characterize the pop-ulation properties but also because a knowledge of thesizes yields albedos and inferences on their surface prop-erties. CCAT will make possible a statistical investiga-tion of both the size distribution and surface propertiesof TNOs. No other instrument or telescope currently un-der development or planned will rival CCAT’s ability tofind these objects.

3. TECHNICAL OVERVIEW

3.1. The FacilityThe CCAT science case, observatory requirements, and

conceptual design were developed as part of a $2M studyjointly funded by Cornell and Caltech/JPL, which re-sulted in a Feasibility/Concept Design Study Report1.

1 http://www.submm.org/doc/2006-01-ccat-feasibility.pdf. Seealso Section 4.

3

The overall specifications for the CCAT observatory arelisted in Table 1. These specifications were derived byconsidering a broad spectrum of science programs rang-ing from studies of Trans-Neptunian objects to surveysfor high z submm galaxies. Commonly recurring themesinclude the need for high sensitivity approaching thatof ALMA, excellent angular resolution to enable targetidentification and overcome spatial confusion, and a widewavelength coverage in order to constrain the spectra andluminosities of dusty objects. These considerations led tothe choice of a 25 m telescope with a 20’ field of view, ahalf-wavefront error (HWFE) around 10µm rms for highaperture e!ciency at 350 µm, and a location on a high(5600 m) mountain site in Atacama above the ALMAplateau for routine access to the 350 µm atmosphericwindow.

Consistently superb observing conditions are crucialfor achieving CCAT’s scientific objectives. For observa-tions at submm wavelengths, a site with very little at-mospheric water vapor is paramount. The proposed sitefor CCAT is at an altitude of 5612 m, on a plateau about50 m below and 200 m east-north-east of the summit ofCerro Chajnantor (Figure 1), within the Science Preserveestablished by the Chilean Government. At 600 m abovethe ALMA site the CCAT site o"ers nearly twice as muchhigh-quality observing time in the crucial 350/450µm at-mospheric windows.

Multiple factors – the 25 m aperture, the improvedatmospheric transparency of the higher site, the use ofbroadband continuum detectors the high aperture e!-ciency resulting from " 10µm rms optics and the possi-bility of achieving sensitivities limited by photon statis-tics and not by receiver noise – combine to give CCATa continuum point-source flux sensitivity in the 350 µmand 450 µm atmospheric windows that is comparable toALMA on a per-pixel basis. (Figure 2 compares the ex-pected CCAT sensitivity to other instruments.) There-fore, with the use of large array cameras, the mappingspeed for CCAT will be many orders of magnitude fasterthan ALMA (potentially 105 for the 100 kilopixel cam-era) enabling large–scale surveys and providing extraor-dinary complementarity with ALMA.

The optical design of CCAT foresees a compactRitchey–Chretien with f/0.4 hyperboloid PM, which al-lows for a relatively compact Callotte-type 40 m dome.as shown in Figure 3. This particular design hasbeen studied in considerable detail by Cana-dian industry and it is likely that this will bethe largest part of the Canadian contribution toCCAT.

3.2. The InstrumentsThe scientific power of CCAT derives from recent ad-

vances in submm detector array technology. Array sizeshave been growing exponentially. At present, the stateof the art is represented by the kilopixel-scale arrays ofsuperconducting transition-edge sensor (TES) bolome-ters now in use at SPT, ACT and in the SCUBA 2 in-strument at the JCMT. For CCAT, the goal is to havecameras with up to " 50 kilopixels available at “firstlight”2. Two cameras are envisioned: one at short wave-lengths, 200#620µm, and one at long wavelengths, cov-ering 740# 2000µm. With 50,000 Nyquist-sampled pix-

els, CCAT’s field of view could be filled at " = 1mm.However, since the number of pixels required scales as"!2, filling the field of view at 350 µm would requireof order 400 kilopixels. Clearly, building instruments atthis scale presents a broad spectrum of technical chal-lenges including cryogenics, optics, ba#ing, shielding,electronics, and mechanical issues. These are long-termchallenges: it is important to remember that CCAT canproduce outstanding science even with existing kilopixel-scale arrays.

Similar challenges exist for spectroscopic instrumen-tation A CCAT goal is to construct broadband direct-detection grating spectrometers, capable of determiningz via the [CII] or CO lines. Existing examples include theZEUS and Z-spec instruments now in use at the CSO.However, CCAT will go beyond these single-beam in-struments and host multi-object spectrometers capableof observing 10–100 objects simultaneously while span-ning multiple atmospheric windows. Again, this presentsnumerous technical challenges. It is now also possible toconstruct large heterodyne array instruments for high-resolution spectral mapping. The 64-pixel 350 GHz sys-tem in construction at U. of Arizona provides one exam-ple; ultimately, heterodyne systems with 256–1024 pixelsshould be possible.

4. MANAGEMENT STATUS

The CCAT partnership was initiated in 2004 throughan MOU signed by Cornell University and the Califor-nia Institute of Technology. The consortium was joinedin 2007 by the University of Colorado, a consortium ofCanadian universities and the United Kingdom ATC.and then in 2009 by a consortium of German Universi-ties and by Associated Universities, Inc. of Washington,D.C.

CCAT will take six years to complete from the datewhen su!cient funding is in hand. Planning has contin-ued to improve as the design has matured and additionalsite information has become available. The current esti-mates of cost for the telescope and associated facilitiesare based on the analysis carried out in the course of theFeasibility Study of 2004–2006. The US CCAT partnersanticipate requesting funds for operations and surveysfrom the NSF. Caltech would close the CSO, presentlysupported by the NSF, in preference for CCAT.

The minimum “buy-in” to be a major partner in CCATis to contribute a 20% share of the costs: $22M (US) plusoperating costs for a minimum of five years. In additionto the dome, Canada will also contribute to instrumentswith contributions very similar to, but a generation moresophisticated than, those built in Canada for the JCMT.These may include parts of the continuum cameras, po-larimeters, and imaging Fourier Transform Spectrome-ters and software both for data analysis and archives. Apotential major all-Canadian contribution is large het-erodyne focal plane arrays (see the white paper by S.Claude!) and the correlator back-ends. All of the pro-posed Canadian contributions are in areas in which thereis a significant heritage and expertise within Canada, andindeed CCAT will help to ensure that the there is an evenhigher return for the investments already made in theseareas. similar to, but a next-generation more sophisti-cated than, those built in Canada for the JCMT.

4

TABLE 1CCAT Specifications

Specification Requirement Goal Remarks

Aperture 25 m sensitivity, confusionWavelength range 350 ! 1400 µm 200 ! 3500 µm dust SEDField of view 10! 20! large-format arraysAngular resolution 3!!.5–14!! 2!!–35!! ! = 1!! " "/(100 µm)

Half Wavefront Error < 12.5 µm < 9.5 µm rms (HWFE)Site conditions < 1mm < 0.7mm median pwvPolarization 0.2% 0.05% after calibrationEmissivity < 10%, " > 300 µm < 5%, " > 800 µm sky loading is low

< 20%, " = 200 µm

Elevation range 5 ! 90" from horizonAzimuth range ±270" from northPointing, blind 2!! 0!!.5 rms

o!set 0!!.3 0!!.2 within 1"

repeatability 0!!.3 0!!.2 rms, one hour

Scan rate 0".2 s#1 1" s#1 slow and fast modesacceleration 0".4 s#2 2" s#2 e"cient scan patternspointing knowledge 0!!.2 0!!.1 rms

Secondary nutation ±2!.5 @ 1Hz, azimuth only

Fig. 1.— Left: Cerro Chajnantor as seen from the ALMA plateau. Right: A view of the CCAT proposed site.

5

Fig. 2.— CCAT continuum point-source sensitivity per pixel compared to other facilities, calculated for a 5# detection in one hourintegration.

Fig. 3.— An artist’s view of CCAT at the proposed site on Cerro Chajnantor